Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Levy, Ido, Shapira, Eilam, Goldshtein, Yinon, Yaeli, Avi, Mashkif, Nir, Shlomov, Segev
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2602.16429
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866908839620640768
author Levy, Ido
Shapira, Eilam
Goldshtein, Yinon
Yaeli, Avi
Mashkif, Nir
Shlomov, Segev
author_facet Levy, Ido
Shapira, Eilam
Goldshtein, Yinon
Yaeli, Avi
Mashkif, Nir
Shlomov, Segev
contents Agentic systems, AI architectures that autonomously execute multi-step workflows to achieve complex goals, are often built using repeated large language model (LLM) calls for closed-set decision tasks such as routing, shortlisting, gating, and verification. While convenient, this design makes deployments slow and expensive due to cumulative latency and token usage. We propose TabAgent, a framework for replacing generative decision components in closed-set selection tasks with a compact textual-tabular classifier trained on execution traces. TabAgent (i) extracts structured schema, state, and dependency features from trajectories (TabSchema), (ii) augments coverage with schema-aligned synthetic supervision (TabSynth), and (iii) scores candidates with a lightweight classifier (TabHead). On the long-horizon AppWorld benchmark, TabAgent maintains task-level success while eliminating shortlist-time LLM calls, reducing latency by approximately 95% and inference cost by 85-91%. Beyond tool shortlisting, TabAgent generalizes to other agentic decision heads, establishing a paradigm for learned discriminative replacements of generative bottlenecks in production agent architectures.
format Preprint
id arxiv_https___arxiv_org_abs_2602_16429
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle TabAgent: A Framework for Replacing Agentic Generative Components with Tabular-Textual Classifiers
Levy, Ido
Shapira, Eilam
Goldshtein, Yinon
Yaeli, Avi
Mashkif, Nir
Shlomov, Segev
Computation and Language
Agentic systems, AI architectures that autonomously execute multi-step workflows to achieve complex goals, are often built using repeated large language model (LLM) calls for closed-set decision tasks such as routing, shortlisting, gating, and verification. While convenient, this design makes deployments slow and expensive due to cumulative latency and token usage. We propose TabAgent, a framework for replacing generative decision components in closed-set selection tasks with a compact textual-tabular classifier trained on execution traces. TabAgent (i) extracts structured schema, state, and dependency features from trajectories (TabSchema), (ii) augments coverage with schema-aligned synthetic supervision (TabSynth), and (iii) scores candidates with a lightweight classifier (TabHead). On the long-horizon AppWorld benchmark, TabAgent maintains task-level success while eliminating shortlist-time LLM calls, reducing latency by approximately 95% and inference cost by 85-91%. Beyond tool shortlisting, TabAgent generalizes to other agentic decision heads, establishing a paradigm for learned discriminative replacements of generative bottlenecks in production agent architectures.
title TabAgent: A Framework for Replacing Agentic Generative Components with Tabular-Textual Classifiers
topic Computation and Language
url https://arxiv.org/abs/2602.16429